SPM: Model Predictive Control -

Model Predictive Control

Contact Details

Dr. Adrian Wills

Email

Phone

+61 2 49216028

Fax

+61 2 49216993

Office

Callaghan Campus
Building EA: EA-204

Post

Dr. Adrian Wills

School of Electrical Engineering and Computer Science

- - -

Funding

Australian Research Council

Discovery Project
DP0774086
2007-2009
Value: $246,090

University of Newcastle

Strategic Pilot Research Grant
Strategic Pilot Research Grant
2008-2009
Value: $15000

University of Newcastle

Near Miss Grant
G0189828
2009
Value: $20000

Australian Research Council

Discovery Project
DP1095645
2010-2012
Value: $170,000

Industry Funding

General Motors Partnership
2011-2012
Value: $110,000

This project is concerned with development of algorithms and hardware for high-speed model predictive control (MPC) solutions. Both linear and nonlinear MPC systems are considered.

Recent Events:

Successfully demonstrated limited precision FPGA design for an on-line MPC controller on the cantilevered beam hardware.
Successfully implemented Non-linear MPC to control an Inverted Pendulum running at 40Hz (Results here)
Hardware design for QP solver successfully matched against the bit-accurate c4Hardware model. Using as few as 6 mantissa bits causes little performance degradation, while allowing for a design that is over 8 times faster than the previous DSP implementation of the same system.

Current Work:

Implementation of NLMPC alogrithm in hardware.
Continue analysis and optimisation of QP Solver design on cantilevered beam hardware

Contents

Objectives

Aims this project.

Background

Background to Model Predictive Control.

Inverted Pendulum

A pilot study using an inverted pendulum apparatus.

Quadratic Programming Solver

Implementation of the Active Set Algorithm in FPGAs and ASICs

Publications

Publications related to this project.

Project Team

Background to the people working on this project.